Terminal Doppler Weather Radar, the Microburst at Dallas-Fort Worth, and the Wind Shear Warning System Built from the Wreckage of Delta One Ninety-One

On August 5, 1985, Delta Air Lines Flight 191 flew into a microburst on final approach to Dallas-Fort Worth International Airport and fell out of the sky. 137 people died. The crash triggered a fundamental transformation in aviation weather detection technology - and exposed coverage gaps that still exist for general aviation pilots today.

What Is a Microburst - and Why Conventional Radar Missed It

A microburst is a small-scale downdraft that, upon reaching the ground, spreads outward in all directions. The typical footprint is two to three miles across. The event usually lasts two to five minutes. Those dimensions sound manageable until you consider what they mean on approach.

Within that footprint, a crew can experience a shift from a 20-knot headwind to a 20-knot tailwind within a single mile - a 40-knot swing in energy. For a transport-category aircraft configured for landing, at low altitude, low airspeed, and reduced power, that shift is potentially unsurvivable.

The headwind gives the aircraft excess energy. The nose rises. The crew pushes over. Then the tailwind strips that energy away instantly. The aircraft is suddenly below approach speed and below approach altitude, with no room to recover.

Conventional radar made the problem worse by making it invisible. Traditional radar measures reflectivity - it shows where precipitation is. A microburst is defined by its wind pattern, not its precipitation intensity. A moderate-looking thunderstorm could produce a killing microburst that appeared unremarkable on the scope.

The Dry Microburst: When Radar Sees Nothing at All

The most dangerous variant is the dry microburst. In arid environments - the desert southwest, the high plains - precipitation can fall from a thunderstorm and evaporate completely before reaching the ground. This phenomenon is called virga: the streaky curtain hanging below a cloud base that never quite arrives.

That evaporation is a cooling process. The air column gets cold, dense, and drops rapidly. When it hits the ground, nothing shows on conventional radar - there is no precipitation to reflect the signal.

That is what killed Delta Flight 191. The crew flew into a storm exhibiting virga. The NTSB report noted radar wasn’t showing significant returns in the area they were transiting. They saw the storm. They discussed it. And then it hit them.

The Accidents That Set the Stage

Delta 191 was not the first microburst disaster. Eastern Air Lines Flight 66 at JFK in June 1975 killed 113 people. Pan American Flight 759 out of New Orleans in July 1982 killed 153 people. The mechanism was roughly the same in each case: a microburst on or near the approach path.

After Eastern 66, the FAA began installing the Low-Level Wind Shear Alert System (LLWAS) - a network of anemometers placed around airports. When wind differences between sensors exceeded a threshold, the tower received an alert. Early installations had as few as six sensors. The system could tell you wind shear existed somewhere in the airport environment. It could not tell you where, at what altitude, or how intense.

Dallas-Fort Worth had that system installed on August 5, 1985. It generated alerts. It wasn’t enough.

How Doppler Radar Solved the Problem

The National Severe Storms Laboratory had already been developing the solution before the Dallas crash. Doppler radar does something conventional radar cannot: it measures the phase shift of the return signal.

When precipitation moves toward the radar, the return frequency is slightly higher. When it moves away, slightly lower. This is the same Doppler effect that changes the pitch of a train horn as it passes.

Apply that to a microburst. Because a microburst spreads outward in all directions after hitting the ground, radar aimed at that cell sees precipitation moving toward it on one side and away on the other - a divergent velocity signature. That pattern is detectable even when reflectivity is low, even when precipitation is light, even when the storm looks unremarkable on a conventional scope.

The Terminal Doppler Weather Radar Program

The FAA launched the Terminal Doppler Weather Radar (TDWR) program in the late 1980s. The first units went operational in the early 1990s. By the mid-1990s, 45 TDWR units were installed at major airports across the country - Atlanta, Chicago, Dallas-Fort Worth, Denver, Houston, Las Vegas, Los Angeles, Miami, Minneapolis, New York, Phoenix, and Seattle among them.

Each unit is a dedicated system, not repurposed from the National Weather Service network. It scans at a low elevation angle optimized for the terminal environment, looking out to roughly 55 nautical miles. The scan cycle is approximately six minutes. A continuous algorithm looks for that divergent velocity signature and immediately alerts the tower when found.

The controller then issues something like: microburst alert, expect a four-zero knot loss at three mile final, one thousand feet, precautionary wind shear advisories in effect.

The results are documented. In the fifteen years before widespread TDWR deployment, microburst-related accidents killed hundreds of people. In the fifteen years after, fatal microburst encounters at airports with radar coverage dropped to essentially zero.

Airborne Wind Shear Detection: The Cockpit Layer

In the early 1990s, the FAA mandated Predictive Wind Shear (PWS) systems on air carrier aircraft. These airborne weather radars apply the same divergent velocity detection principle - from the cockpit looking forward on the approach path.

Current systems can detect a microburst approximately 20 to 45 seconds before the aircraft reaches it. That window is enough to initiate a go-around before entering the shear.

This created meaningful redundancy: ground-based radar alerting the tower, airborne radar alerting the crew directly. If the ground system missed something, the aircraft radar might catch it. If the aircraft radar was ambiguous, the ground alert reinforced it.

Where the System Still Has Gaps

The 45 TDWR units cover major airports. They do not cover every commercial airport. They do not cover general aviation airports.

If you operate at a regional field or a general aviation airport, there is no Terminal Doppler Weather Radar watching your approach path. The upgraded Low-Level Wind Shear Alert System serves some of these airports, but coverage is not universal.

The dry microburst problem also persists. TDWR relies on Doppler velocity measurements of precipitation. A genuinely dry microburst with minimal precipitation reaching radar-detectable levels remains harder to catch than a wet one. Dual-polarization upgrades are improving this, but the gap has not closed.

General aviation pilots also lack airborne PWS capability. Most panel-mount weather radar is reflectivity-only - it shows where the rain is, not what the wind inside it is doing.

What General Aviation Pilots Actually Have

PIREPs (Pilot Weather Reports) remain the most time-sensitive, location-specific wind shear data available. A pilot who flew the approach ten minutes ahead and reported severe turbulence or a significant airspeed excursion is delivering information no radar has synthesized for you. File them. Read them. They are the most underused tool in the weather briefing stack.

Convective SIGMETs are issued when thunderstorms produce surface winds of 40 knots or greater, hail of 3/4 inch or greater, or embedded thunderstorm activity. If convective SIGMETs are active on your route, microburst risk is real and present.

The 20 nautical mile rule from active convective cells is not a conservative buffer. It is based on documented outflow radii from major cells.

FIS-B weather through the ADS-B infrastructure delivers NEXRAD composite imagery, METARs, TAFs, and advisories to equipped aircraft at no subscription cost. The critical caveat: FIS-B NEXRAD imagery can be 5 to 15 minutes old. A cell that looked manageable 15 minutes ago may have produced an outflow boundary now sitting directly across your approach corridor. FIS-B is a situational awareness tool - not a real-time radar picture.

The go-around is always available. The NTSB has noted consistently that wind shear accidents share a common element: pilots who pressed the approach when conditions were deteriorating.

Where the Technology Is Heading

The National Weather Service dual-polarization upgrades to the NEXRAD network improve the ability to detect virga signatures, differentiate rain from hail, and filter ground clutter - meaningful progress on the dry microburst detection problem. The FAA is working on algorithmic improvements to leverage this data in TDWR units as they are upgraded.

Some of those 45 TDWR units are aging. Whether to recapitalize the network or integrate wind shear detection into a future consolidated radar architecture is an active policy discussion.

The National Severe Storms Laboratory is researching rapid-scan Doppler methods that could shorten the update cycle from six minutes to one or two minutes. That matters: microbursts evolve fast, and a six-minute-old picture on an active storm day is useful but not current.

Key Takeaways

The 1985 crash of Delta Flight 191 at Dallas-Fort Worth, which killed 137 people, directly drove the development and nationwide deployment of the Terminal Doppler Weather Radar program.
Microbursts are defined by wind pattern, not precipitation intensity - making them invisible to conventional reflectivity-only radar, especially in dry virga conditions.
45 TDWR units at major airports, combined with mandatory airborne Predictive Wind Shear systems on air carriers, reduced fatal microburst encounters at covered airports to essentially zero.
General aviation pilots have no TDWR coverage at most airports they fly from, no airborne PWS capability, and must rely on PIREPs, convective SIGMETs, visual discipline, and the go-around.
FIS-B NEXRAD carries 5 to 15 minutes of latency - it shows where weather was, not where it is right now.